Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
The temporal and spatial dynamics of prosodic processing - an MEG study with French listeners
Poster E81 in Poster Session E, Thursday, October 26, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
This poster is part of the Sandbox Series.
Chantal Oderbolz1, Martin Meyer1,2; 1University of Zurich, 2University of Klagenfurt, Austria
In speech, any given target utterance can be realized through many different physical manifestations that depend on factors like linguistic intentions or speaker-related characteristics. At the same, it is well-known that listeners use prosody – acoustic-phonetic modulations – to make linguistic inferences, for example, about an utterance’s lexical segmentation or syntactic structure. This suggests that prosody is an important part of the linguistic system and functions as a useful source of information for speech processing. However, these observations raise the question of how listeners navigate phonetic variability to extract a dependable and generalizable cue. One frequent source of phonetic variability stems from sentence types – for example, statements and polar questions – and the distinctions between them. The prosodic feature responsible for conveying this distinction is fundamental frequency (f0). Polar questions are reliably produced with a final rise in terminal f0 while statements are marked by a falling terminal f0. For speech perception to be efficient and successful in such a case, a listener must, on the one hand, perform a fine-grained analysis of the physical prosodic contour and, on the other hand, form an abstract prosodic representation that allows a categorization regarding sentence type. A broad network, especially in the right hemisphere, has been implicated in the analysis and transformation of prosodic information including the superior temporal gyrus (STG) and sulcus (STS), inferior frontal gyrus, as well as premotor cortex. However, what is missing from these mainly fMRI-based findings is a characterization of how the brain captures meaningful phonetic variability and the temporal and spatial dynamics of associated processing stages. To fill this gap, we are currently running an MEG study with French-speaking young adults. Participants listen to and categorize French words intoned either as a question or a statement. To create controlled phonetic variability, we synthesized a continuum between two prototypes by manipulating f0 in 6 equidistant steps. The psychometric identification function stemming from a behavioral pilot study confirms that these stimuli are perceived categorically. On a neural level, we expect to see an initially continuous physical representation of phonetic variability in early auditory processing regions especially in right posterior STS. Progressing in space (towards anterior STS) and time, we expect to find a categorical representation of prosodic cues. Given previous research, we hypothesize the premotor cortex to be involved in the transformation from continuous to categorical representations. From a broader perspective, this study is valuable insofar as prosody is generally an understudied phenomenon in the cognitive neuroscience of speech: Studies usually control for prosodic cues as an undesired confounding source. This is surprising considering the fundamental role prosody plays in language acquisition in infants as well as speech processing in adults. Therefore, the insights generated from this experiment will contribute to a more encompassing understanding of speech processing in the brain. Additionally, since comparisons between the human brain and more complex computational models of language are getting increased attention, an integration of prosodic information into these architectures will better capture the naturalistic aspects of how we communicate.
Topic Areas: Speech Perception, Prosody